Opinion
B
ack in November, I did an
interview with a good friend and
industry colleague Antony Reeve-
Crook on the subject of what can we
achieve with matchmaking technology,
and what some of the challenges
are (featured on The Exhibitionists
podcast).
The very word ‘matchmaking’ has
become a hot topic in the exhibition
industry, and somewhat of a buzzword.
Sadly, the widespread use of the
term has led to a narrowing of the
concept. Where we should be building
recommendation engines that can
recommend different things (or
people) to a user, we instead assume
that everyone has the same objective
of filling up their diary with as many
meetings as possible.
We’re not there yet
Matchmaking technology can do great
things, but it is far from realising its full
potential, even from the perspective
of already available tech. If we look
outside of our industry to giants such as
Amazon and Netflix, who have all built
powerful recommendation engines,
we can glimpse some of the potentials
for such technology. For example,
over 80 per cent of watched titles
on Netflix come from the personal
recommendation list, with an average
daily stream time of two hours per user.
Meanwhile, 23 per cent of Amazon
visitors are there simply for inspiration
and are not looking to buy a specific
product. These are impressive statistics
64 — March
Matchmaking:
what’s going on?
EN interviews Dr Mykyta
Fastovets, CTO, at
ExpoPlatform about what’s
happening in matchmaking
technology
in terms of both engagement and
prediction accuracy. These results set
goals that we can only strive towards in
our own industry.
So, how can we achieve similar
results, and what obstacles are in the
way? There are two key areas that need
to be improved in order to increase the
efficiency of matchmaking solutions:
onboarding, and engagement.
The onboarding problem
Onboarding is a key element of
ensuring end-users (in our case it’s
visitors and exhibitors) are informed
of the capabilities of the tools at their
disposal, are educated on how to use
these tools, and are provided effortless
access to them. Crazy though it may
seem, many organisers have not found
the right way to communicate with
visitors and exhibitors.
Some time ago I had a meeting with
a client where the issue of onboarding
came up. It transpired that the
organiser knew that exhibitors don’t
read their emails yet no other means
of communication was possible. Why
aren’t we experimenting more with
what communication mediums actually
work?
The other issue that often comes up
is around accounts and passwords.
Everyone has so many accounts and so
many passwords to remember that it’s
very easy to lose users before they’ve
even logged in. Simple solutions like
auto-login links, and sticky sessions can
go a long way in ensuring a frictionless
transition from an onboarding email to
logging into your account. Yet, we still
often come across randomly generated
passwords, one-time passcodes, and
complicated integration schemes where
the actual matchmaking system is
hidden away by layers of menus, pages,
and most importantly, unnecessary
additional effort.
Engagement and relevance
It should be obvious that in order
for a system to be engaging it must
cater towards the user’s needs and
expectations in terms of ease of use,
but also in terms of the relevance
of recommendations. If the system
is not intuitive and if there are no
instructions, even the most motivated
users will abandon their attempts to
engage. Equally, even with the best
possible design and user experience,
if the recommended results are
consistently not relevant users will
quickly lose interest and abandon the
system. The challenge is that each
person has a different definition of
what is relevant.
Of course, even if we do understand
the objectives of a user, we still need
data to make accurate prediction;.
data that is consistently labelled,
standardised, and accurate.
Moving forward
We’ve had some very promising
experiences working with person-to-
object recommendations and there
is no reason why we can’t provide an
equivalent to the Netflix or Amazon
experience for digital touchpoints for
exhibitions. To make true progress
a few things need to happen. First,
organisers need to take control of their
data. Second, we need to abandon the
notion that we know what our visitors
and exhibitors want better than they
do. Finally, we need to experiment
more with AI algorithms and try
different approaches.
Dr Mykyta
Fastovets CTO at
ExpoPlatform